it looks like you have the correct design. I would look for an
interaction between group and age (contrast 0 0 1 -1) in the dods model.
If there is none, then go to a doss model with 1 -1 0 contrast. If there
is an interaction, then some say you cannot test for a simple difference
in the groups
Dear Freesufers,
I am currently running mri_glmfit using a relatively simple design: Two
groups and one covariate. Design is: 1 -1 0 0 .. to control for the
co-variate (age).
It is the example with two groups and one covariate in your web-page
http://freesurfer.net/fswiki/Fsgdf2G1V
I was under t
